A Machine Learning Model Predicts Crop-Energy Yields in Agrivoltaics

Agrivoltaic (AV) systems can simultaneously produce food and energy, providing a much-needed solution to balance the food-energy nexus for the growing global population. In this context, it is crucial to identify optimal design solutions applicable to worldwide locations with reliable food and energ...

Full description

Saved in:
Bibliographic Details
Main Authors: Mahmood, Farrukh ibne, Alam, Muhammad Ashraful, Butt, Nauman Zafar
Format: Conference Proceeding
Language:English
Subjects:
Agrivoltaics
crop yield
Crops
Data models
energy estimation
Estimation
Machine learning
Neural Networks
Photovoltaic systems
Photovoltaics
Predictive models
Reliability engineering
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Agrivoltaic (AV) systems can simultaneously produce food and energy, providing a much-needed solution to balance the food-energy nexus for the growing global population. In this context, it is crucial to identify optimal design solutions applicable to worldwide locations with reliable food and energy output. Typically, physics-based or mechanistic models are used to predict the food-energy yield for AV. However, these models require significant processing power and time, particularly when applied to multiple locations. Therefore, in this paper, we employ machine learning (ML) to predict crop and photovoltaic (PV) yields for an AV system. Our data is derived from an in-house physics-based model validated with field measurements. We demonstrate how the ML-based model can rapidly (within minutes) and accurately (R 2 ~ 0.99) predict yields for thousands of locations, a task that would take days using the physics-based model.
ISSN:2995-1755
DOI:10.1109/PVSC57443.2024.10749352